Jules Quentin KOUAMO

Abstract: Nowadays, integrated systems have digital and analog parts cohabiting on the same chip (or in a close environment). These two parts, which are intimately linked in their operation, have disjointed tests that evolve in completely different worlds. This traditional separation is no longer justified in technological terms: in modern circuits, digital and analog electronics coexist and merge, giving rise to complex mixed-signal systems in which the two worlds are closely linked. Unfortunately, the skills, techniques and general paradigms are still extremely different in the two worlds: while ad hoc solutions to specific problems exist, the systematic consideration of the specific constraints of both domains for a unified and coherent solution is not sufficiently addressed. This thesis aims at proposing a methodology and an infrastructure that will allow the development of hybrid software and hardware approaches optimized to the needs of the designers. Mixed-signal systems must go through multiple phases of design, validation, and testing, and each has its own tools and constraints that make sharing information and porting solutions nearly impossible. One of the objectives of the thesis is to contribute to bridge the technological and methodological 'gap' between these stages and to evaluate the feasibility of a unified approach, with a particular focus on the life cycle.
The first occupation of the PhD student will be to acquire skills in both digital and analog test domains by interacting with the two supervisors whose complementary skills cover well both. On the digital side, the focus will be on functional testing through the IEEE 1687 standard and its derivatives. The candidate will have to master the complete flow from test specification to hardware/software co-simulation and prototyping. On the analog side, will have to understand the functional and structural testing techniques of mixed analog circuits. The candidate will be expected to master the various facets of mixed-signal testing, with particular emphasis on approaches that require interactive algorithms, but do not have strict timing requirements, and thus can be run remotely.
Once this learning phase is complete, the candidate will be expected to propose a unified approach: using digital techniques to drive analog tests in an efficient and portable manner. This step will require the specification, implementation, and development of extensions and patches to current workflows and tools that have never been used in this framework. It will need to specify and develop techniques for routing analog data through the container.